Annotation of dietlibc/CHANGES, revision 1.14
1.9 fefe 1: 0.7.1:
1.11 fefe 2: optimized away the static arrays from __ltostr and __lltostr.
3: optimized away the static arrays (and more!) from strtoul and strtoull.
4: added system (by Olaf).
1.12 fefe 5: added tolower and toupper
1.13 fefe 6: added getopt from Olaf (~600 bytes on x86!)
1.14 ! fefe 7: added clone for mips (by Olaf).
! 8: fixed longjmp and setjmp on sparc not to require glibc include files.
1.9 fefe 9:
10: 0.7:
1.6 olaf 11: clone support for i386,alpha,arm and sparc by Olaf.
1.1 cvs 12: Initial dynamic loading glue for i386, also by Olaf.
13: We don't have a dynamic loader yet, so this is not yet usable.
14: added readv, sigdelset, sigaltstack, sigfillset, sigismember, usleep,
15: vsprintf, writev, all by Olaf.
1.2 fefe 16: imported readdir fix from Guillaume. Mhh, can the kernel interface
17: really be this broken?
1.3 fefe 18: Found a kludgy way to detect whether the kernel includes define struct
19: stat64 (they also define STAT64_HAS_BROKEN_ST_INO). If this does
20: not work for you, please tell me!
1.4 fefe 21: Removed bogus __seek_types enum, it's a bunch of #defines now.
1.7 fefe 22: Fixed stdio line buffering. The code was already there but I forgot
23: to mark stdio and stdout as line buffered. This should really be
24: done dynamically using isatty, though.
1.8 fefe 25: Removed superfluous "if (1)" in execvp.c
1.1 cvs 26:
27: 0.6.14:
28: *scanf did not append the 0 byte for %s.
29: added creat and changed creat to call open directly instead of open64.
30: did for fread what 0.6.10 did for fwrite.
31: repaired buffered stdio a little (line buffering is still not
32: supported).
33: [All of these bugs have been reported by Guillaume. Thanks!]
34: added __pure__ to strchr and strrchr in <string.h>
35:
36: 0.6.13:
37: included COPYING to make clear that the diet libc is covered by the
38: GNU General Public License (and _not_ the LGPL). That means that
39: you need to obtain a license from me if you want to use the diet
40: libc in a proprietary program that you want to distribute.
41: The x86 unified syscall swapped arguments #4 and #5 and noone
42: noticed until now! Thanks to Guillaume Cottenceau for reporting
43: this!
44: The malloc had a subtle bug with small allocations that could cause
45: segfault. Guillaume reported it, Olaf fixed it.
46:
47: 0.6.12:
48: added assert, statfs and fstatfs
49: added include/net/if.h so fget compiles again.
50:
51: 0.6.11:
52: added putenv.
53:
54: 0.6.10:
55: fixed fwrite. It returned the number of bytes written, not the number
56: of records. Thanks to Albert D. Cahalan for pointing this out.
57: fixed fgets. It returned EOF on empty lines. Thanks, Erik Frey.
58: added brk and sbrk. harold@nb.com.sg made me do it.
59:
60: 0.6.9:
61: fixed execvp not to return on ENOENT
62: did for {open|read|close|seek|tell}dir what I did for stat before.
63: The glibc compatibility part is currently non-functional.
64: fixed white space handling in sscanf " (" would not match " (".
65: fixed printf to accept 'l' flag.
66: fixed getpwuid (ignored last line)
67:
68: 0.6.8:
69: Olaf contributed initial thread-safe syscalls for i386, alpha, sparc,
70: mips and arm. I did the one for ppc. They can be enabled in dietfeatures.h.
71: added strlcpy and strlcat from OpenBSD.
72: added stat64, fstat64 and lstat64.
73: added endian.h and made ending of strlen.c endianness-aware
74: added string routines size tweak to dietfeatures.h. Most of the
75: string routines have been unrolled. This is up to three times as
76: fast but creates up to three times larger code. Now you can disable
77: the unrolling.
78: added sendfile.
79: I also added pread. It works on x86 and sparc but not on ppc and mips.
80: I have no idea what's going on. strace disagrees with the kernel.
81: Can anyone help? I asked the Linux kernel mailing list for help, too.
82: added a few aliases of the type __libc_open for open. nm on
83: libpthreads.so indicates that we will need them.
84:
85: 0.6.7:
86: the sources now compile without warnings with -Wall.
87: printf now prints "(null)" when %s is passed NULL. This can be
88: removed with WANT_NULL_PRINTF in dietfeatures.h
89: added vfprintf, execl, ttyname.
90: stat, lstat and fstat are now #defined to __dietstat, __dietlstat and
91: __dietfstat so they can use the normal kernel struct stat.
92: stat, lstat and fstat are now C wrappers that convert to the glibc
93: struct stat. So we can avoid that overhead for programs that use
94: dietlibc headers.
95:
96: 0.6.6:
97: changed the Makefiles so you can now set CFLAGS on the command line
98: added memccmp (analogous to memccpy) and strncmp.
99: Olaf fixed another bug in the sparc unified syscall.
100: Paul Clifford contributed a C version of his strlen.S that is much
101: more efficient than the previous strlen on all platforms! It also
102: looks like technology from Roswell. ;-}
103: added a "real" stdio. Well, almost. Please contribute!
104: added dietfeatures.h so you can remove features you don't need.
105: removed debug code from vsnprintf that changed \0 to ' ' (argh!).
106: stdio now works with simple programs and minigzip from libz.
107: stdio uses some major trickery to avoid linking the stdio, stderr and
108: stdout (and reserving space for them) when they are not used.
109:
110: 0.6.5:
111: Red Plait <redplait@ixcelerator.com> found several bugs in diet libc,
112: most of them bugs in the new header files, but also missing sigset
113: functions.
114: Paul Clifford contributed an assembly strlen.S for ARM and fixed
115: several bugs.
116: Fixed ppc/setjmp and mips/pipe.
117: Olaf contributed a new sparc unified syscall.
118:
119: 0.6.4:
120: declared is* static inline in getservent.
121: added assert_fail, strtoul, isalpha, isdigit, isalnum, isascii.
122: changed strlen to return 0 when passed NULL.
123: new, much smaller unified syscall for MIPS. Thanks to Olaf the Mad
124: Scientist who actually implemented this without access to a MIPS
125: box, just from reading the architecture manual. And it worked out
126: of the box.
127: started a set of system includes, mainly so that I can use lcc and
128: my alpha-linux cross compiler (which is unable to cross-compile
129: glibc). diet libc can now be compiled without any include files
130: from a normal libc. The includes do declare more than diet libc
131: currently delivers and they are still far from complete for real
132: applications.
133: I will only add assembly versions that are smaller _and_ faster than
134: the C version. Larger routines are only accepted if they are called
135: very often and are substantially faster.
136: added i386 assembly strchr, which is smaller and faster than the old
137: version. It is, however, larger than the version contributed by
138: proton (thanks, anyway).
139: added i386 strlen (31% faster, 14% smaller)
140: "make t" will now create a map file called "mapfile".
141:
142: 0.6.3:
143: added sys_errlist, strerror and perror
144: added isblank
145: added atol
146:
147: 0.6.2:
148: mmap for ARM didn't compile. Thanks, Paul!
149:
150: 0.6.1:
151: split mmap into the architecture specific subdirectories.
152: getenv now copes with environ==NULL, thanks Paul Clifford.
153: Paul also contributed a smaller ARM startup code.
154:
155: 0.6:
156: strcat returned the wrong result. Thanks, Dietz Pröpper.
157: strtod now understands a negative exponent (oops, thanks Bertram Barth)
158: Port to arm-linux-gnu, but on the Netwinder I use for testing the
159: __dtostr does not work (I have no idea why!)
160: The MIPS port now uses (much smaller) non-PIC code. That means that
161: the applications you link against diet libc also have to be compiled
162: non-PIC. I suggest copying the CFLAGS from the diet libc Makefile.
163: Thanks to Ralf Bächle for helping me with this!
164: I made subdirectories for the architectures and use VPATH to override
165: VPATH so that make finds the source file automatically. That should
166: simplify the sources greatly.
167: Olaf Dreesen contributed Alpha support including setjmp and longjmp!
168: ARM and MIPS now also have setjmp and longjmp
169: "compile" and "load" are now make targets. Use them for djb programs.
170: added contrib/elftrunc.c which will remove unnecessary ELF headers.
171: Again, contributed from Olaf. Great work!
172:
173: 0.5.12:
174: printf also does signed numbers.
175: If you don't use atexit, dietlibc now does not link exit, only _exit.
176: *printf now correctly returns the number of bytes written.
177: Olaf contributed experimental sscanf and vsscanf implementations.
178: If passed NULL as buffer, snprintf will not write anything but still
179: return the number of bytes it would have written.
180: Initial MIPS port! (Oh, the agony!)
181: No setjmp and longjmp support yet! Please contribute!
182: I even implemented unified syscalls for MIPS. Still, MIPS code is
183: almost twice the size of SPARC code. If anyone knows why: please
184: tell me!
185:
186: 0.5.11:
187: I implemented new unified syscalls for x86, sparc and ppc, this time based
188: on .s files and not .c files, and I moved the syscalls into
189: subdirectores to clean the dietlibc sources up a little.
190: A binary consisting of printf("%s is %d\n","olaf",23) is now
191: 2864 bytes on sparc
192: 2488 bytes on intel
193: I kludgily implemented fprintf(stdout,... and fprintf(stderr,...
194: to make a few more applications work.
195: Olaf Dreesen also implemented some unified syscall stuff for x86.
196: He discovered that with our unified syscall interface it actually
197: costs <10 bytes total to make all system calls thread safe!
198: I implemented atexit() (can register up to 4 callbacks).
199: I implemented strtod and __dtostr (the opposite). Now we can add
200: floating point support to vsnprintf and sscanf!
201:
202: 0.5.10:
203: I actually saw that I can not only merge the errno handling code of
204: the system calls on x86, I can also merge the rest (including the
205: arguments) except for setting the system call number. All those
206: system calls are now just a jump to a unified system call handler.
207: I got rid of x86openclose again. The savings are substantial:
208: chown from embutils went from 7664 to 7184 bytes!
209: If I'd move the system call wrappers to assembly language, I could
210: even reduce the alignment (does not matter since it's just a jump
211: anyway) and get rid of the "ret" behind each jump (one byte per
212: system call!)
213:
214: 0.5.9:
215: added memchr, strpbrk, strstr, strtol, isspace (hehe)
216: fixed strdup
217: fixed return values for strcpy and strcat (thanks to Norbert Berzen)
218: Olaf Dreesen contributed a strtol and initial {sn|vsn|}printf implementation
219: (no signed integers, only strings and unsigned integers (but
220: supporting octal, hex and decimal). Thanks, Olaf! (by the way:
221: that code must be wonderful, I don't understand it at all *bg*)
222: Olaf's printf does understand padding like in "%08d" and "%8d" and
223: automatically pads pointers with '0'.
224: Rewrote the x86 start code, old: 22 instructions, new: 12 instructions.
225: Thanks to proton for inspiration on this.
226: The new x86 start code also does not reference exit any more, thus saving
227: 64 bytes for executables that don't call exit explicitly.
228: [insert maniacal laughter] EVERY BYTE COUNTS!1!! ;-)
229: Moved errno and environ to start.S, saving no byte binary size but
230: speeding up compilation and slightly shortening dietlibc.a ;-)
231: For x86: joined open and close into one assembler file, sharing the
232: errno handling. This is a feasibility test and it actually saves a
233: few bytes. I think I will reimplement the _syscall[1-6] macros on
234: all platforms now to share the errno handling code for them.
235:
236: 0.5.8:
237: fixed strchr to be able to look for 0.
238:
239: added _llseek, ftruncate, getpgid, getresgid, getresuid, getsid,
240: memccpy, memmove, mprotect, setregid, setresgid, setresuid, setreuid,
241: strncpy, swapon, truncate, strtok, strtok_r, strspn, strcspn (all
242: contributed by Olaf Dreesen)
243:
244: added execvp, getcwd
245:
246: fixed __xmknod
247:
248: Note: can it be that the ftw interface really is so broken that I
249: cannot implement it without having to implement some searching data
250: structure? I included an experimental ftw implementation that will
251: not follow symlinks.
252:
253: 0.5.7:
254: oops, if_nametoindex was broken!
255:
256: 0.5.6:
257: ported to ppc-linux.
258: fixed i386 sigsetjmp (I mistyped the function name)
259: included sigjmp.c
260:
261: 0.5.5:
262: added wait, sys_siglist, longjmp/setjmp/sigsetjmp for i386 and sparc
263: actually, it wasn't fork that wasn't working for sparc, it was pipe.
264: I fixed it now.
265: dietlibc/SPARC assumes -msupersparc (does not provide div, mul, etc)
266: fixed readdir to use getdents and not the intel inline asm
267:
268: 0.5.4:
269: ported to sparclinux (sparc32 only). Beware: does not work yet.
270: added raise, abort, readlink, strcat, geteuid, geteuid, wait3, access
271:
272: 0.5.3:
273: added tcsetattr and getenv to compile e3
274:
275: 0.5.2:
276: added vhangup, tcgetattr, isatty and memcmp for fgetty
277: (http://www.fefe.de/fgetty/)
278: added localtime from uC-libc.
279:
280: 0.5.1:
281: stat and friends actually work now. Yuck! Another case of
282: translation between kernel and userland. Why can't the kernel
283: people and the libc people simply agree on a standard?
284: removed many unnecessary includes to speed up compilation.
285: added -fomit-frame-pointer and i386 compilation target to reduce code size.
286: inlined socketcall to reduce code size.
287: "load" and "compile" are examples for djb code.
288:
289: 0.5:
290: Olaf Dreesen contributed a much smaller implementation of malloc and friends.
291: each object file is now treated with "strip -x -R .note -R .comment"
292: split each system call into a separate object file
293: added ntohs, htons, alarm, if_indextoname and if_nametoindex
294:
295: 0.4.1:
296: oops, the strchr implementation was wrong. Thanks Jens Laas!
297:
298: 0.4:
299: copied opendir and friends from uC-libc.
300: added getservent and getservby* in gerservent.c
301:
302: 0.3:
303: remove readdir system call
304: added getdents system call
305: added getpwnam/getpwuid implementation in getpwnam.c
306: added a few string functions (in str*.c)
307:
LinuxTV legacy CVS <linuxtv.org/cvs>